Change in E-cigarette Risk Perception and Smoking Behavior of Black and Latinx Individuals who Smoke

Abstract

Introduction

Trends in knowledge and beliefs about e-cigarette (EC) harm have been changing. Our study examined whether change in risk perception of ECs predicts cigarettes smoked per week (CPW) among Black and Latinx individuals who smoke enrolled in an EC switching randomized clinical trial (RCT).

Methods

We analyzed data from one arm of a 6-week EC RCT of individuals who smoke attempting to switch to nicotine salt pod system ECs (n=110; Black, n=57; Latinx, n=53). Our explanatory variable was change in risk perception of ECs compared to combustible cigarettes (CC) from baseline to week 6. Our outcome was CPW measured by a 7-day timeline follow-back interview. A negative binomial GEE model was conducted to examine the association between risk perception and CPW at baseline and week 6.

Results

The mean CPW decreased from 82.8 (SD=49.8) at baseline to 15.8 (SD=29.8) at week 6. A one-level increase in EC risk perception (i.e., EC perceived as riskier than CC from baseline to week 6) was associated with an increase in CPW (IRR: 1.83, 95% CI: 1.03; 3.24). Latinx participants were more likely to have higher CPW as EC risk perception increased compared to Black participants (IRR=1.89, 95% CI: 1.09; 3.26).

Conclusion

We found that CPW at week 6 was significantly higher as change in risk perception of EC relative to CC increased. Given the influence of risk perception on smoking behavior, people who smoke should be educated on the benefits and harms of ECs.

1. Introduction

Electronic cigarettes (ECs) are battery-powered devices that produce an inhalable aerosol by heating a liquid solution of humectants (usually propylene glycol or glycerol), nicotine, and flavorings.¹ ECs entered the United States (U.S.) market in 2007 and began to gain widespread popularity in 2014.² Companies initially marketed ECs as harm reduction substitutes for individuals who use combustible cigarettes (CC)³ as they reduced exposure to toxicants produced through combustion.⁴ Additionally, ECs were seen as a device that could help individuals who use CCs quit. A recent meta-analysis found moderate evidence that ECs with nicotine increased quit rates for at least six months and resulted in higher cessation rates compared to ECs without nicotine and nicotine replacement therapy.⁵ However, the 2020 Surgeon General’s Report on smoking cessation stated there is inadequate evidence on the efficacy of ECs for smoking cessation among adults and they are not approved by the Federal Drug Administration (FDA) for this purpose.⁶

From a harm reduction perspective, complete substitution of ECs for CCs, as reported by the National Academies of Sciences, Engineering, and Medicine, reduces exposure to toxic combustion gases and particles for people who smoke CCs.⁴ Short-term evidence has shown that the levels of toxicants such as formaldehyde, acetaldehyde, acrolein and toluene⁷ are significantly lower in ECs compared to CCs, supporting the idea that substitution of ECs for CCs may reduce exposure to tobacco-specific toxicants.⁸ Hence, completely switching from CCs to ECs will likely result in reduced short-term adverse health outcomes.⁵ A recent randomized clinical trial (RCT) looked at the effect of fourth generation pod-based ECs on carcinogen exposure among Black and Latinx who smoke.⁹ This study found that individuals who smoke and were randomized to receive 4th generation ECs (5% nicotine) for 6 weeks were more likely to quit CCs, smoke fewer CCs in the past seven days, and have lower levels of biomarkers of tobacco exposure 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) and carbon monoxide (CO) at 6-week follow-up compared to those randomized to the control condition of continued CC smoking.⁹ At week 6 of this study, 28.1% of participants in the EC group exclusively smoked ECs, 57.9% used both ECs and CCs, and 14.0% used CCs exclusively.⁹

One factor that plays a role in switching and product use decision for people who smoke CCs is the initial perception of the harmfulness of ECs compared to CCs. Prior studies suggest low to moderate EC knowledge^10–13 and beliefs that ECs are healthier and more socially acceptable than CCs among predominately White people who smoke CCs and nonsmokers.¹⁰ One review found that adult and youth who use ECs, CCs, both ECs and CCs, and those who use neither of these products perceived ECs as healthier, safer, and less addictive than CCs.¹⁴ Additionally, predominantly White adults who use both ECs and CCs who perceive ECs as less harmful were more likely to switch to exclusive EC use.¹⁵ However, EC risk perceptions are changing. While some studies prior to 2014 showed that ECs were perceived as less harmful than CCs, following studies have found ECs are perceived as equally or more harmful than CCs among adults who smoke CCs, individuals who don’t smoke CCs and ECs, and youth who smoke CCs.¹⁴

There is limited research on EC risk perception by race/ethnicity. One study found that Black and Hispanic participants who smoke were more likely to believe ECs were more harmful than CCs and less likely to begin using ECs in the context of smoking cessation compared to White participants who smoke.¹⁶ Similarly, another study found that Black participants had lower EC health risk perceptions (i.e. ECs are less harmful to health than CCs), compared to both White and Hispanic participants who smoke.¹⁷ Harm perceptions of ECs have largely been studied in predominately White people who smoke and/or predominately White youth and young adults.

Few studies have examined risk perception differences between racially and ethnically minoritized groups and many are limited in demonstrating switching behaviors. One study looking at primarily White U.S. adults who use both ECs and CCs found that those who deemed ECs as “less harmful” than CCs were more likely to switch to exclusive EC use, more likely to continue smoking both ECs and CCs, and less likely to switch to exclusive CC use one year later compared to individuals who smoke both ECs and CCs who had an EC risk perception of “about the same,” “more harmful,” or “don’t know”.¹⁵ Similar studies should be conducted for Black and Latinx individuals who smoke.

Given the disproportionate burden of smoking-related death and disease among minoritized groups compounded with the low rates of switching from CCs to ECs among these populations,^16,18 it is important to study switching patterns among Black and Latinx individuals who smoke CCs. Furthermore, with increases in targeted tobacco advertising in racially and ethnically minoritized neighborhoods,¹⁹ it is important that we gain insight into the current perceptions these populations have about ECs to better tailor public health interventions and research to specific population groups with the ultimate goal of smoking cessation.

The primary objective of this study was 1) to examine whether change in risk perception of ECs predicts the number of cigarettes smoked per week (CPW) at the end of a 6-week RCT. The secondary objectives of this study were 2) to examine how race/ethnicity moderates the relationship between change in risk perception of ECs and CPW, and 3) to compare change in risk perception of EC use between Black and Latinx participants. We hypothesize that 1) people whose risk perception of ECs increases will have higher CPW at week 6 compared to those whose risk perception of ECs decreases, 2) compared to Latinx individuals, Black individuals will be less likely to experience an increase in CPW as EC risk perceptions increase, because they are more likely to have overall lower EC risk perception and lower rates of CPW, and 3) given that prior literature showed Black individuals perceived ECs to have less health risks than White and Latinx individuals,¹⁷ we hypothesize that Black individuals who smoke will be more likely to decrease their EC risk perceptions compared to Latinx individuals.

2. Methods

2.1. Data

This study is a secondary data analysis of Project Switch, an unblinded RCT of participants interested in switching from CCs to ECs. Participants were randomized in a 2:1 ratio to receive 6 weeks of JUUL ECs (5% nicotine) or continue smoking their usual brand of cigarettes.⁹ Participants in the EC arm also received education, training, and action planning for a complete switch to ECs.

Recruitment began in May 2018 and primary follow-up was completed in May 2019. Study procedures were approved and monitored by the California State University San Marcos Institutional Review Board and the University of Kansas School of Medicine Institutional Review Board. All participants provided written informed consent.

2.2. Participants

Participant eligibility and exclusion criteria, study methods and procedures were described in detail in the main study.⁸ Briefly, Hispanic/Latinx participants (N=94) were recruited from San Diego, California, and Black participants (N=92) from Kansas City, Missouri. Individuals who self-identified as Latinx were classified as such, regardless of race. Participants were eligible if they were 21 years or older, smoked at least five cigarettes per day on at least 25 of the past 30 days, smoked for at least six months, had an expired carbon monoxide reading of greater than five parts per million (PPM) at baseline, and had a systolic blood pressure less than 160 mm Hg and diastolic less than 105 mm Hg at baseline.

Participants were excluded if they primarily used other tobacco products or used equal amounts of other tobacco products and CCs, used ECs on four or more of the past 30 days, were actively enrolled in a smoking cessation program or other clinical trial, and used smoking cessation pharmacotherapy in the past 30 days. Participants were also excluded for health concerns including hospitalization for mental illness in the past 30 days, heart-related illness in the past 30 days, and unstable mental status or health status. Those who were pregnant, breastfeeding, or planning to become pregnant in the next 6 months were also excluded. Finally, participants were excluded if they resided with another person enrolled in the study or were planning to move away from San Diego or Kansas City during the study period.

All participants were fluent in English or Spanish and had regular access to a telephone and were willing to switch from smoking CCs to ECs for 6 weeks. This study examined longitudinal outcomes among 110 participants randomized to the EC group who provided data at the final week-6 visit.

2.3. Measures

2.3.1. Change in risk perception of e-cigarette use

Risk perception of ECs was assessed at baseline using the survey question “Is using e-cigarettes less harmful, about the same, or more harmful than smoking cigarettes?” Responses were “less harmful,” “about the same,” “more harmful,” or “I don’t know.”²⁰ Responses were recategorized into three response categories: 1 = “less harmful,” 2 = “about the same or I don’t know,” and 3 = “more harmful”.

Change scores were calculated for risk perception of ECs by subtracting baseline risk perception from week 6 risk perception. A negative change score indicates a decrease in risk perception of ECs from baseline to week 6 (i.e., EC less harmful than CC), a positive change score indicates an increase in risk perception of ECs from baseline to week 6 (i.e., EC more harmful than CC), and 0 indicates no change in risk perception from baseline to week 6.

2.3.2. Cigarettes smoked per week (CPW)

The outcome of this study was CPW measured by 7-day TLFB^21,22 at week 6. 7-day TLFB was used because it assesses small differences in CPW that may not impact CO. CO alone was not considered as an outcome due to between-person variation in CO levels with similar changes in cigarette consumption.

2.4. Covariates

Demographic characteristics were measured at baseline. Covariates included age (years), gender (male; female), race/ethnicity (Black; Latinx), education level (less than a high school diploma/GED; high school diploma/GED or more), federal poverty level (above 200% of the federal poverty level; 200% federal poverty level and lower), marital status (never married; other) and self-reported overall health (good; less than good).

2.5. Statistical Analyses

Demographic, health, and smoking characteristics at baseline and week 6 were compared using regressions and chi-square tests. Mean change in risk perception was also estimated according to baseline participant characteristics and compared using t-tests. The frequency of change in risk perception was estimated by race/ethnicity as well as by the frequency of EC substitution at week 6 using chi-square tests.

A negative binomial GEE model was used to test the effect of the change in risk perception on CPW, adjusting for all covariates. A second model was estimated with a two-way interaction between change in risk perception and number of study weeks to test differences in the effect at baseline and week 6. A third model was estimated with the two-way interaction between change scores and race/ethnicity to test differences on the effect of change in risk perception on CPW by race/ethnicity. Models two and three were also adjusted by study covariates.

3. Results

Participants were a mean age of 44.4 years (SD=12.8) and over half the sample were male (59.1%). 51.8% identified as Black and 48.2% as Latinx. The mean CPW decreased from 82.8 (SD = 49.8) at baseline to 15.8 (SD = 29.8) at week 6. At baseline, 28.2% of participants thought ECs were “less harmful” than CCs, while at week 6, the majority of participants responded that ECs were “less harmful”. At baseline, all participants exclusively smoked CCs. Complete participant characteristics are presented in Table 1.

Table 1.

Participant characteristics at baseline and week 6 (N=110)

	Baseline N=110	Week 6 N=110
Variables	M (SD) or n (%)	M (SD) or n (%)
Mean number of cigarettes	82.84 (49.79)	15.83
smoked per week (SD)*		(29.83)
Risk Perception of ECa vs. CCb***
Less harmful	31 (28.18)	67 (60.91)
About the same or I don’t know	74 (67.27)	42 (38.18)
More harmful	5 (4.55)	1 (0.91)
Age	44.4 (12.75)	-
Gender
Male	65 (59.09)	-
Female	45 (40.91)	-
Race/Ethnicity
Black	57 (51.82)	-
Hispanic/Latinx	53 (48.18)	-
Education Status
High school diploma/GED or more	97 (88.18)	-
Less than a high school diploma/GED	13 (11.82)	-
Federal Poverty Level
At or below 200%	81 (73.64)	-
Above 200%	29 (26.36)	-
Marital Status
Never married	51 (46.36)	-
Other	59 (53.64)	-
Self-Reported Overall Health***
Less than good	43 (39.09)	31 (28.18)
Good, very good or excellent	67 (60.91)	79 (71.82)

^aEC=e-cigarettes

^bCC=combustible cigarettes

P-value comparing baseline to week 6:

^*<0.05;

^**<0.01;

^***<0.001

Mean change in EC risk perception according to baseline participant characteristics is presented in Supplementary Table 1. Mean change in risk perception did not differ by study characteristics, including race/ethnicity.

Negative binomial GEE results are displayed in Table 2. Overall, the number of CPW significantly decreased by week 6 of the study (IRR=0.19, 95% CI: 0.13; 0.28). There were no differences in CPW at baseline by race/ethnicity. Change in EC risk perception was not associated with baseline CPW. However, change in EC risk perception was positively associated with CPW at week 6. Every one-level increase of EC risk perception was associated with an increase in CPW at week 6 (IRR: 1.83, 95% CI: 1.03; 3.24). The results from Figure 1 show that the predicted mean CPW for those with a change in EC risk perception score of −2 significantly decreased from 69.3 CPW at baseline to 4.0 CPW at week 6. Those with a change in EC risk perception score of −1 significantly decreased mean CPW from 78.4 to 8.2. Those with a change in EC risk perception score of 0, that is, those who did not change their EC risk perception, significantly decreased their CPW from 88.7 at baseline to 16.9 at week 6. Those with an increase in EC risk perception with a score of 1 also had a significant decrease in CPW from baseline to week 6 (100.4 compared to 35.0 respectively).

Table 2.

Negative binomial regression of risk perception and CPW^a at week 6 (N=110)

Variables	IRR (95% CI)
Change in Risk Perception of ECb vs. CCc	1.13 (0.90; 1.43)
Weeks	0.19 (0.13; 0.28)***
Change in Risk Perception*Weeks	1.83 (1.03; 3.24)*
Age	1.03 (1.01; 1.04)***
Gender (ref “Male”)
Female	0.94 (0.69; 1.27)
Race/Ethnicity (ref “Black”)
Hispanic/Latinx	1.43 (0.95; 2.16)
Education Status (ref “high school diploma/GED or more”)
Less than a high school diploma/GED	1.79 (1.01; 3.17)*
Federal Poverty Level (ref “above 200%”)
At or below 200%	1.12 (0.78; 1.62)
Marital Status (ref “never married”)
Other	0.67 (0.48; 0.94)*
Self-reported overall health (ref “Good, very good or excellent”)
Less than good	0.86 (0.60; 1.23)

^*<0.05;

^**<0.01;

^***<0.001

^aCPW=cigarettes smoked per week measured by 7-day timeline follow-back interview

^bEC=e-cigarettes

^cCC=combustible cigarettes

Figure 1.

Average marginal effects of Model 2 testing the interaction between EC risk perception and study week, adjusted for covariates

Race/ethnicity moderated the relationship between EC risk perceptions and CPW. Latinx participants were more likely to have higher week 6 CPW as EC risk perception increased compared to Black participants (IRR=1.89, 95% CI: 1.09; 3.26) (Figure 2). The results from Figure 2 show that for those with no change in risk perception (score of 0), Black individuals had a mean CPW of 28.6 and Latinx individuals had a mean CPW of 52.9 (p=0.03). There were no significant differences in CPW by race/ethnicity for the other EC risk perception scores.

Figure 2.

Average marginal effects of Model 3 testing the interaction between EC risk perception and race/ethnicity, adjusted for covariates

*CPW is significantly different for individuals with a change in EC risk perception of 0 (p=0.03). There were no significant differences in CPW by race/ethnicity for the other EC risk perception levels (−2, −1, 1).

4. Discussion

Our first hypothesis stated that individuals whose risk perception of ECs increases will have higher CPW at week 6 compared to those whose risk perception of ECs decreases. Our findings confirmed our hypothesis as they showed that CPW at week 6 was significantly higher as risk perception scores increased. While participants with increased EC risk perception did have lower CPW at week 6, the magnitude of this difference is narrower compared to those whose EC risk perception decreased (change score −1 or −2) or did not change (change score 0). Previous studies have shown that among individuals who use both ECs and CCs, those who perceive ECs as less harmful were more likely to switch to exclusive EC use.¹⁵ Our findings show a similar trend as those whose EC risk perception decreased were more likely to have lower CPW than those whose EC risk perception increased.

Intervention influences may have also contributed to the overall decrease in CPW from baseline to week 6. The intervention component included a discussion about the barriers and benefits of switching to ECs were discussed with the participant and action planning for exclusive switching was revisited.⁸ The change in risk perception from baseline to week 6 suggests that the intervention may have had an impact on participants’ risk perceptions, and the overall decline in CPW.

Our second and third hypotheses on racial/ethnic differences in CPW based on change in risk perception were also confirmed. While there was no difference in EC risk perception by race/ethnicity, the effect of change in risk perception on CPW varied by race/ethnicity. As hypothesized, there was a significant association such that for Black individuals, CPW was relatively constant for all change scores. However, for Latinx individuals, there was a significant increase in CPW as EC risk perception increased. These findings point to a potential socio-cultural difference in risk perception of ECs. The observed difference in CPW between Black and Latinx individuals may be due to overall lower risk perception of ECs among Black individuals. While there are few studies on EC risk perceptions and smoking behavior among racially and ethnically minoritized individuals, there is evidence that Black individuals, and even Black adolescents, held lower EC health risk perceptions (i.e. ECs are not as bad as CCs for your health) than Latinx individuals.^17,23 If Black participants hold lower initial EC risk perceptions, it is possible that their CPW across EC risk perception change scores remained constant whereas those of the Latinx participants increased. Consequently, the association between risk perception and CPW was not as pronounced for Black individuals. Yet, a study found that Black individuals were more likely to use ECs as a cessation aid compared to Hispanics.²⁴

Our study has several limitations that should be considered. First, EC risk perception was assessed with a single item, for which the validity has not been verified with other questionnaire items. Additionally, the EC risk perception responses were recategorized to combine the answer choices of “about the same,” and “I don’t know.” It is possible that separating these answer choices may yield different results with other implications. Additionally, given that the study was over a 6-week period, it is not certain whether this was enough time for participants to commit to completely switching and if the switching behavior would be maintained over longer periods of time. Furthermore, this sample size was small, so our findings should be replicated with a larger sample size. However, this study was powered to detect small to moderate effects. Since our sample is specific to Black participants from the Midwest and Latinx from the West Coast, it may not generalize to other Black and Latinx individuals in other parts of the country.

5. Conclusion

This secondary data analysis found that CPW at week 6 was significantly higher as change in EC risk perception scores increased. While CPW declined overall, the magnitude of decline was smaller for those whose EC risk perception increased than those whose EC risk perception remained constant or declined. Race/ethnicity was a moderator of the change in risk perception of ECs and CPW such that CPW was constant across EC risk perception change scores for Black individuals but increased for Latinx individuals. Given the impact risk perception has on CPW, it is important to understand how various components of the intervention such as education, training and action planning changed participants’ risk perception of ECs. Future research should examine the impact of education on forming risk perceptions among Black and Latinx people who smoke and consider how to effectively disseminate knowledge of ECs as a potential harm reduction tool.

Highlights.

• Decreased e-cigarette risk perception associated with fewer cigarettes/week

• Differences in e-cigarette risk perception between Latinx and Black participants

• Risk perceptions of e-cigarettes can influence smoking behavior